ArcBest

Less-than-truckload (LTL) carriers operate on thin margins, and city delivery routes — the final leg from freight terminal to customer — are among the hardest to optimize. Each day, ABF Freight dispatchers at terminals across the country managed dozens of stops per driver, with variable shipment weights, tight time windows, and fluctuating daily load density. Inefficient sequencing forced ABF to absorb excess mileage, higher diesel consumption, and reliance on purchased transportation and rented equipment to handle overflow demand. These external city cartage costs — paying third parties to complete deliveries ABF's own fleet couldn't efficiently cover — compounded at scale and were a persistent drag on operating margins.

ArcBest developed an AI-based city route optimization tool in-house for ABF Freight, applying reinforcement learning and combinatorial optimization to dynamically sequence daily city delivery routes. Rather than relying on static dispatch plans, the system continuously recalculates optimal stop sequences across ABF's terminal network using real-time and historical shipment data, minimizing total mileage and fuel consumption per route. No third-party vendor is identified — this was a proprietary internal build, integrated directly into ABF Freight's existing dispatch operations. After validating results on the delivery side, ArcBest launched a follow-on pilot to apply the same optimization framework to the shipment pickup process — treating inbound routing as a natural extension of the same underlying model rather than a separate project.

The initiative delivered more than $1 million per month in operating income improvement, a figure CEO Judy McReynolds disclosed directly to investors on ArcBest's Q4 2023 earnings call. The program contributed to measurable margin recovery at ABF Freight:

• $64.3M Q4 2023 operating income — a 28% year-over-year increase from $50.2M in Q4 2022
• Reduced purchased transportation and rented equipment costs
• Lowered external city cartage expenses across the terminal network

ABF Freight President Seth Runser described the tool as part of a broader optimization roadmap, with additional initiatives actively in development.

• In-house AI development can deliver investor-grade results: ArcBest built this tool internally and reported impact at the monthly P&L level, demonstrating that proprietary optimization is viable without a third-party vendor.
• City delivery routing is a high-ROI entry point for AI in LTL freight: the problem is well-scoped, dispatch data already exists, and savings compound across every terminal daily.
• Delivery-to-pickup expansion doubles the addressable opportunity: applying the same optimization logic to inbound pickup routing is a low-friction second phase that avoids rebuilding from scratch.
• Mileage reduction is a multi-lever win: fewer miles simultaneously cuts fuel costs, purchased transportation spend, and carbon footprint — strengthening the business case across finance, operations, and sustainability.